Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

<@ULVA73B9P> hi we added variables for prefect-server    - name: PREFECT_SERVER_DATABASE_SQLALCHEMY_POOL_SIZE
      value: 100
    - name: PREFECT_SERVER_DATABASE_SQLALCHEMY_MAX_OVERFLOW
      value: 20
    - name: PREFECT_SERVER_DATABASE_SQLALCHEMY_POOL_TIMEOUT
      value: 10  for avoid loop and timeout error , but cuurenly get timeout exec_wal error

:arrows_counterclockwise: Researching your answer... this may take a while


:wrench: Using tool: `research_prefect_topic`

:bar_chart: Tools used:
  • `check_cli_command` (1x)
  • `display_callable_signature` (2x)
  • `explore_module_offerings` (3x)
  • `read_github_issues` (1x)
  • `research_prefect_topic` (2x)
  • `search_github_discussions` (1x)

<@U09BH7ZKPD2> can you provide more of the error message you're seeing? Are you using Postgres or SQLite for your database?

I use Postrgres aws rds, prefect server 3.4.8,  timeout error appear under load when many tasks run

```    tr, pr = await connector
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/site-packages/asyncpg/connect_utils.py", line 931, in __connect_addr
	2025-08-26 17:15:01.280	
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
	2025-08-26 17:15:01.280	
    return await __connect_addr(params, True, *args)
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/site-packages/asyncpg/connect_utils.py", line 886, in _connect_addr
	2025-08-26 17:15:01.280	
           ^^^^^^^^^^^^^^^^^^^^
	2025-08-26 17:15:01.280	
    conn = await _connect_addr(
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/site-packages/asyncpg/connect_utils.py", line 1049, in _connect
	2025-08-26 17:15:01.280	
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
	2025-08-26 17:15:01.280	
    return await connect_utils._connect(
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/site-packages/asyncpg/connection.py", line 2421, in connect
	2025-08-26 17:15:01.280	
Traceback (most recent call last):
	2025-08-26 17:15:01.280	
15:15:01.245 | ERROR   | prefect.server.services.failexpiredpauses - Unexpected error in: TimeoutError()
	2025-08-26 17:15:01.280	
TimeoutError
	2025-08-26 17:15:01.280	
    raise TimeoutError from exc_val
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/asyncio/timeouts.py", line 115, in __aexit__




	2025-08-26 17:15:01.280	
               ^^^^^^^^^^^^^^^^^^^^^^^
	2025-08-26 17:15:01.280	
    async with compat.timeout(timeout):
	2025-08-26 17:15:01.280	
  File "/usr/local/lib/python3.12/site-packages/asyncpg/connection.py", line 2420, in connect
	2025-08-26 17:15:01.280	
            ^^^^^^^^^^^^```

actually  it's aurora, we suppose it should  be supported  it's db.t4g.medium

Screenshot 2025-08-27 at 13.58.52.png

I also saw heavy write load during load testing

Yes, write load is expected as we'll be inserting metadata about flows/tasks into the database.

The AWS `t` series is a *burstable* instance type. This means you get a certain amount of CPU credits on an hourly basis. Once you exceed your available credits, your instance will be throttled. This is most likely why you're seeing slowness after a period of sustained DB writes.

I would recommend you look at using something like  `db.c6gd.large` for your RDS instance type. This will give you the same CPU and memory resources as the `db.t4g.medium` instance type, but the `c` series instance does not burst and you receive a higher performance baseline.

ok, got it, thanks  it's dev instance on prod we have more powerful instance